Trends in hip fracture rates in US hemodialysis patients, 1993-2010.

ABSTRACT

BACKGROUND: Changes in mineral and bone disorder treatment patterns and demographic changes in the dialysis population may have influenced hip fracture rates in US dialysis patients in 1993-2010. STUDY DESIGN: Retrospective follow-up study analyzing trends over time in hospitalized hip fracture rates. SETTING & PARTICIPANTS: Using Medicare data, we created 2 point-prevalent study cohorts for each study year. Hemodialysis cohorts included patients with Medicare as primary payer receiving hemodialysis in the United States on January 1 of each year; non-end-stage renal disease (ESRD) cohorts included Medicare beneficiaries 66 years or older on January 1 of each year. FACTORS Age, sex, race, primary cause of ESRD, dual Medicare/Medicaid enrollment status, comorbid conditions. OUTCOMES: Hip fracture rates. MEASUREMENTS Unadjusted hip fracture rates measured using number of events per 1,000 person-years in each year, then adjusted for patient characteristics. Poisson models estimated strata-specific event rates. RESULTS: The observed number of first hospitalized hip fracture events and the adjusted hip fracture rate increased steadily from 1993 (831 events; 11.9/1,000 person-years), peaked in 2004 (3,256 events; 21.9/1,000 person-years), and decreased through 2010 (2,912 events; 16.6/1,000 person-years). The trend for the subset of hemodialysis patients 66 years or older was similar to the trend for the full hemodialysis cohort; however, it differed markedly in magnitude and pattern from the non-ESRD Medicare cohort, for which rates were substantially lower and slowly decreasing since 1996. LIMITATIONS: Unable to provide causal explanations for observed changes; hip fractures identified through inpatient episodes; results do not describe hemodialysis patients without Medicare Parts A and B; laboratory values unavailable in the Medicare data set. CONCLUSIONS: Temporal trends in hip fracture rates among Medicare hemodialysis patients differ markedly from the steadily decreasing trend in non-ESRD Medicare beneficiaries, showing a relatively rapid increase until 2004 and relatively rapid decrease thereafter. Further research is needed to define associated factors.